BMS-CTMC-2025-461Ubiquitin as a Putative Causal Immune Orchestrator in Initial COVID-19 Severity: A Mendelian...

The initial COVID-19 outbreak precipitated unprecedented health crises, with research impeded by containment protocols and the high mortality rate. While SARS-CoV-2 perturbs immune ubiquitin dynamics, the causal interplay between them remains uncharacterized.

We integrated multi-omics and Mendelian randomization (MR) approaches, including bidirectional and two-step MR, to assess causality between ubiquitination and initial COVID-19 severity and to determine whether immune cells mediate this relationship. Transcriptomic profiling, protein-protein interaction networks, immune-correlation analysis, and an elastic-net-derived prognostic model were employed to explore these interactions further and identify potential biomarkers.

Bidirectional MR confirmed causal links between COVID-19 severity and ubiquitination, identifying 13 ubiquitin-related genes with altered expression. A two-step MR analysis of 731 immune cell traits revealed CD45RA+ CD8+ T cells as a mediator of severe COVID-19-induced ubiquitin dysregulation (mediation proportion: 5.797%). A prognostic model based on ISG15, UBE2B, and UBE2V1 demonstrated high predictive accuracy in the training cohort (AUC up to 0.977) and maintained good predictive power in the testing cohort (AUC: 0.672-0.819), with significant risk stratification (log-rank P < 0.001).

We propose that ubiquitin may function as a causal immune orchestrator, shaping clinical outcomes in the initial phase of COVID-19. The derived ubiquitin gene prognostic model demonstrates potential for clinical risk stratification, while pathway enrichment and druginteraction analyses suggest opportunities for repurposing ubiquitin-related targets in COVID-19 therapy.

This ubiquitin-centric framework advances precision management of severe COVID- 19, highlights the potential of ubiquitin-related targets for drug development, and establishes a multi-omics MR-based paradigm for future viral pandemics.